[Mplayer-users] new audio code in CVS
atmosfear at users.sourceforge.net
Thu May 10 09:35:01 CEST 2001
On Thursday, 10. May 2001 05:56, you wrote:
> I've just commited the new audio code of mplayer.
> - audio playback rewritten, using OSS specs (should be more standard)
> - does not use select(), but uses ioctl()'s
> (SNDCTL_DSP_GETOSPACE,GETODELAY) => much precise timing!!! (if card&driver
> supports it...)
> - PTS correction is now in the demuxer (see ds_tell_pts())
> - input bitrate (bytes/sec) is calculated & stored in sh_audio_t struct
> - option '-alsa' no more used, it always uses usleep()
> Now for most precise timing it summarizes all delays:
> - delay in demuxer buffers (handled by ds packets)
> - pts correction (ds_tell_pts())
> - compressed bytes in the a_in_buffer (ACM/DShow)
> - uncompressed bytes in the a_buffer (everything, except PCM+aLaw)
> - buffered bytes in the audio card's/DMA buffer. (ioctl GETODELAY)
> I got <10ms error for most mpg, avi and asf files, it's very good.
> But VOB seems to be broken, at least it's about +-200ms :((( I'll fix it.
> Seeking not tested, and not finished, do not report bugs about it.
> I'll finish it very soon.
> Frame dropping disabled!
> (it will be implemented later, this hack didn't work for most users)
> I'm interested in results, so is it better or worse than before,
> are your audio problems solved etc. Please all the time include mplayer -v
> output, it prints many new information about audio card and driver!
> Do not report problems about seeking (including A-V sync problems _after_
> seek), as I said above, it isn't finished yet.
On my ESS Alegro it plays in sloooooooooooooowmotion and no statusbar.
maybe 3spf or so :)
> A'rpi / Astral & ESP-team
> mailto:arpi at thot.banki.hu
> Mplayer-users mailing list
> Mplayer-users at lists.sourceforge.net
Mplayer-users mailing list
Mplayer-users at lists.sourceforge.net
More information about the MPlayer-users